Caltech’s High Contrast High-Resolution Spectroscopy for Segmented telescopes Testbed (HCST)D. Mawet, J.K. Wallace, J. Wang, E. Choquet, J. Fucik, G. Singh, G. Ruane, S. Wieman, and R. Dekany

High contrast imaging on the ground

• Ground-based 8-meter class telescopes: • SPHERE & GPI large surveys halfway through (current yield on the low side) • SCExAO reaching science operations • P3K/P1640 nearing end of life (Gene Serabyn’s SDC still in operations) • LBTAO returning great image quality, struggling with DSM/weather issues • MagAO king in visAO Hα niche • First-generation facilities winding down with a few exceptions (L-band niche

at Keck-NIRC2, VLT-NACO, LMIRCAM at LBT, all equipped with state-of-the-art vortex coronagraphs)

• Extremely Large Telescopes under construction (sort of): • First generation instrument unlikely to be optimized high contrast imagers

(some interesting capabilities with TMT-IRIS and ELT-METIS though). • R&D for second-generation planet finder starting now

High contrast imaging in space

• HST now: archival data and disks, not very competitive for planets (mostly due to limited telescope time access)

• JWST 2018: great for transit spectroscopy. Inner working angle will be >0”.5 (NIRCAM & MIRI)

• WFIRST 2025: excellent contrast, very low throughput, yield and spectroscopic capabilities will be modest (tech demo)

• LUVOIR/HDST/HabEx 2035: currently being defined, opportunity for synergistic developments (this meeting is a good start!)

Most future telescopes will be large and segmented (HabEx?)

• Need to develop coronagraph and wavefront technologies for large segmented telescopes

• Many theoretical designs exist (APLC, PIAACMC, RAVC, etc.)

• Very few lab demonstrations of these concepts so far (none?)

• # HCI facilities in the US dedicated segmented telescopes: 2

• HiCAT at STScI, focussing on space-based projects

• Keck telescope !!!

Coronagraphy on segmented telescope now: Keck NIRC2 L-band vortex

1 λ/D

1 λ/D

Absil, Bottom, Campbell, Carlogmano, Choquet, Delacroix, Femenia, Gomez, Huby, Jolivet, Karlsson, Matthews, Mawet, Reggiani, Serabyn, Wertz, Wizinowich

Large telescopes enable high resolution spectroscopy

• Interesting concepts merging HCI and HRS have been proposed (Snellen et al. 2015)

• And demonstrated!

• The future of high contrast imaging of planets is high contrast high resolution spectroscopy

YES there are enough photons!

• Not trying to measure/trace individual lines.

• The line profile, or cross-correlation peak combines the information of 1000s of molecular lines theoretically resolved at high spectral resolution

• (think about how RV gets to <1m/s, with 1-10km/s resolution per line)

noisy highres data

highres noiseless theoretical template

cross-correlation peak

speckle noise mostly irrelevant, part of the continuum

Renyu Hu (JPL)

Cross correlation signal extraction from noisy data

Ji Wang (Caltech)

Science and technology demonstrator at Keck

Ji Wang (Caltech)

Cross-correlate with spectral templates of molecular species

Konopacky et al. 2013

Measure planet spin

LETTERdoi:10.1038/nature13253

Fast spin of the young extrasolar planet b Pictoris bIgnas A. G. Snellen1, Bernhard R. Brandl1, Remco J. de Kok1,2, Matteo Brogi1, Jayne Birkby1 & Henriette Schwarz1

The spin of a planet arises from the accretion of angular momentumduring its formation1–3, but the details of this process are still unclear.In the Solar System, the equatorial rotation velocities and, conse-quently, spin angular momenta of most of the planets increase withplanetary mass4; the exceptions to this trend are Mercury and Venus,which, since formation, have significantly spun down because of tidalinteractions5,6. Here we report near-infrared spectroscopic observa-tions, at a resolving power of 100,000, of the young extrasolar gasgiant planet b Pictoris b (refs 7, 8). The absorption signal from carbonmonoxide in the planet’s thermal spectrum is found to be blueshiftedwith respect to that from the parent star by approximately 15 kilo-metres per second, consistent with a circular orbit9. The combinedline profile exhibits a rotational broadening of about 25 kilometresper second, meaning that b Pictoris b spins significantly faster thanany planet in the Solar System, in line with the extrapolation of theknown trend in spin velocity with planet mass.

Near-infrared, high-dispersion spectroscopy has been used to char-acterize the atmospheres of hot Jupiters in close-in orbits10,11. Suchobservations use changes in the radial component of the orbital velocityof the planet (resulting in changes in Doppler shift) to filter out the quasi-stationary telluric and stellar contributions in the spectra. Here we makeuse of the spatial separation and the difference in spectral signaturebetween the planet and star, which allow the starlight to be filtered out.A similar technique12 has been applied very successfully13 at a mediumspectral dispersion to characterize the exoplanet HR8799c. We observedthe bPictoris system7,8 (apparent magnitude, K 5 3.5) using the Cryo-genic High-Resolution Infrared Echelle Spectrograph14 (CRIRES) locatedat the Nasmyth focus of Unit Telescope 1 of the Very Large Telescope (VLT)of the European Southern Observatory (ESO) at Cerro Paranal in Chile onthe night of 17 December 2013, with the slit orientated in such way that itencompassed the planet and star.

An important step in the data analysis is the optimal removal of thestellar contribution along the slit, which for this class-A star consists mostlyof a telluric absorption spectrum. The resulting spectra were cross-correlatedwith theoretical spectral templates constructed in a similar way as in ourprevious work on hot Jupiters10,11, varying the planet’s atmospheric tem-perature pressure (T/p) profile and the abundances of carbon monoxide,water vapour and methane, which can also show features in the observedwavelength range. We note that there is a strong degeneracy between theatmospheric T/p profile and the abundance of the molecular species,meaning that different combinations of these parameters result in nearlyidentical template spectra.

At the expected planet position, a broad, blueshifted signal is apparent(Fig. 1), which is strongest when the cross-correlation is performed with aspectral template from an atmospheric model with deep carbon monoxidelines and a small contribution from water. We estimate the signal to have asignal-to-noise ratio of 6.4 by cross-correlating the residual spectrum witha broadened model template, and compare the peak of the cross-correlationprofile with the standard deviation. If we use the cross-correlation profileas seen in Fig. 1 to estimate the signal-to-noise ratio, we need to take intoaccount the width of the signal and the dependence of adjacent pixels inthe profile. This results in a signal-to-noise ratio of 7.8, but we found thelatter method to be less accurate because it does not properly includecontributions from correlated noise structures on scales of the broadsignal. Cross-correlation with the optimal spectrum of water vapour aloneprovides a marginal signal at a signal-to-noise ratio of ,2, which meanswe cannot claim a firm detection of water in the planet’s atmosphere. Nosignal is retrieved for methane models (Extended Data Fig. 1).

We fit the planet profile using a grid of artificial cross-correlation func-tions, produced by cross-correlating the optimal template spectrum with abroadened and velocity-shifted copy of itself, for a range of velocities androtational broadening functions. The best fit (Fig. 2a) was obtained for

1Leiden Observatory, Leiden University, Postbus 9513, 2300 RA Leiden, The Netherlands. 2Netherlands Institute for Space Research (SRON), Sorbonnelaan 2, 3584 CA Utrecht, The Netherlands.

1.0

Rel

ativ

e po

sitio

n (″

)

CC

sig

nal

0.5

0.0

–0.5

–1.0

–60 –40 –20Velocity (km s–1) Velocity (km s–1)

0 20

Planet position

Star positionS

tella

r vel

ocity

40 –100 –50 0 50

a b

Figure 1 | Broadened cross-correlation signal of b Pictoris b. a, CO 1 H2Ocross-correlation signal (linear colour scale) as function of the position alongthe slit (orientated 30u east of north), after the stellar contribution was removed.The x axis shows the radial velocity with respect to the system velocity(120 6 0.7 km s21) of the star15. The y axis denotes the relative position withrespect to the star bPictoris with the planet located 0.499 below, both indicated

by horizontal dashed lines. A broad signal, at a signal-to-noise ratio of 6.4 isvisible, blueshifted by 15.4 6 1.7 km s21 (1s) with respect to the parent star.b, Cross-correlation (CC) signal at the planet position. The dotted curve showsthe arbitrarily scaled auto-correlation function of the l/Dl 5 100,000(resolving power) model template, indicating the CC signal expected from anon-rotating planet.

1 M A Y 2 0 1 4 | V O L 5 0 9 | N A T U R E | 6 3

Macmillan Publishers Limited. All rights reserved©2014

Snellen et al. 2014

Doppler imaging => maps

Luhman 16 (Crossfield et al. 2014)

Caltech HCST goals

• Simulate realistic segmented apertures: • Keck • TMT • LUVOIR/HDST (HabEx?)

• (Simulate atmospheric turbulence) • Includes adaptive optics/wavefront controller, with a dynamic

wavefront sensor and options for amplitude control (2DM EFC, ACAD, etc)

• Includes classical 3-plane single stage coronagraph (apodizer, focal-plane mask, Lyot stop)

• Includes a back-end instrument • Goal: high-res fiber-fed diffraction limited low-noise spectrograph

Top-level requirements

• Realistic telescope simulator • match F number of Keck & TMT first • simulate segment cophasing errors

• Wavelength range: 0.6 to 1.8 microns • Keck / TMT: WFS at Y or J, Science at H / K • Space-based telescope: 0.6 to 1 microns • IWA ~ 1-2 λ/D, OWA ~ 15 λ/D (superNyquist possible) • Accommodate various transmissive coronagraph designs • Minimize Talbot effects (design philosophy similar to HiCAT) • Contrast goals (average over dark hole):

• Raw, phase control only (static, no turbulence): 1e-5 • Raw, phase & amplitude (static, no turbulence): 1e-7 • After HRS: 1e-8, actual limits to be explored

• Inject starlight and faint planet light (with distinct spectral signature) => unique feature

Latest HCST design

J. Fucik (Caltech) / J.K. Wallace (JPL) Many thanks to HiCAT team (E. Choquet, M. N’Diaye, R. Soummer

HRSCoronagraph

LOWFS

AO/WFC

Telescope simulator

Telescope simulator based on IRIS AO segmented DM

37Segments(Keck) 313Segments(TMT)

NOTE:LatestTMTprimarymirrordesignhas492segments

Coupling AO / coronagraph to IR HRS

MRI: Development of a Near-IR Precision Radial Velocity Facility at the W. M. Keck Observatory

January 13, 2016

-9-

Figure 9: The NIST self-referenced EOM laser comb spans a full octave covering the JHK astronomical bands. The purple line represents a simulation while the black curve shows laboratory data. A simplified block diagram is shown at the top as discussed in the text.

Figure 8: Instrumental layout for the self-referenced, broadband laser comb

the ~18 dB insertion loss of the modulators. The result is a deterministic ~24 nm bandwidth frequency-comb spectrum with International System of Units (SI) traceable mode spacing. The comb is then filtered with the aid of a low-finesse optical cavity filter that reduces noise from conversion of electronic thermal noise to comb modes. A programmable line-by-line optical amplitude/phase filter then ‘shapes’ the EOM-comb spectrum to obtain the highest peak-power pulses. These pulses are delivered to the spectral-broadening section which consists of a high-power optical amplifier (EDFA) and a length of “Dispersion-Decreasing Highly NonLinear Fiber” (DD-HNLF). The DD-HNLF for the NIRSPEC system will be constructed by splicing lengths of commercially-available HNLF with discrete dispersions which can control the evolution of the comb spectrum, separately optimizing the short and long wavelength regions. The final step is to use the broadened EOM comb light both for f-2f locking to provide absolute stabilization of the CW laser frequency

and to send the comb signal to NIRSPEC via ~10 m lengths of H or K single mode fiber.

The available comb power is 20 to 50 dB greater than the femtoWatt levels required per comb tooth, so that the spectrum can, to a certain extent, be tuned for improved spectral flatness. If necessary, NIRSPEC’s own filters can be used to isolate particular parts of the spectrum, e.g. the CO bandhead at 2.3 Pm, if steep gradients remain in comb envelope (Figure 9).

3.c.2.2. Feeding NIRSPEC from the AO System

Two optical subsystems feed star light and laser comb light into NIRSPEC. The first assembly (Figure 10) focuses starlight from the AO system into single mode fibers while the second formats the outputs of the fibers into the slit plane of NIRSPEC (Figure 11). Both of these assemblies as well as the K band fiber optic cabling are being developed for an approved Heising-Simon Foundation grant for $700k (D. Mawet, PI) to enable non-PRV, K-band spectroscopy of exoplanets using the AO system. We describe this facility briefly and emphasize that it is being designed with the PRV application in

Figure 10: Light from the AO system is injected into optical fibers on the AO bench. There are 2 fibers in H-band and two in K-band so that one fiber looks at the star while the other monitors the sky for beam-switching between the two positions. This fully-funded addition to the existing AO bench includes accomodation for relay optics and counter-rotating prisms for atmospheric dispersion correction, pickoff of J band light for active tracking, and coupling the f/15 beam into the f/3 fibers.

J.K. Wallace (JPL)

Keck AO to NIRSPEC Fiber Injection Unit concept

HCST status

• Big items acquired:

• 1 BMC 32x32 MEMS DM (2nd DM pending funding)

• Supercontinuum source + fibers

• HASO WFS

• IR and visible cameras

• All optics specified, purchase request is out

• Wish list: OCAM2K, CRed-one (First light imaging), second BMC DM

• Dream list: Large format IR-APD arrays (2Kx2K), two 2K BMC DMs

Short term goals & priorities

• Support to privately funded Keck fiber injection unit (FIU) project: link from Keck AO to NIRSPEC to do high-resolution spectroscopy of known young giant planets

• Support for Keck FIU coronagraph APRA proposal:

• Several phenomenon not scalable wrt wavelength cannot be reproduced in the lab (segment gaps) and needs the real full scale demonstration => Keck

• Support demonstration of coronagraph concepts for future segmented telescopes: TMT, LUVOIR/HDST, HabEx